Parameter-Efficient LLM Unlearning through Sparse Attention Head Updates

Published in Korean Institute of Industrial Engineers, (KIIE), 2025